Rheumatoid arthritis is associated with locoregional decalcification, which can be clearly demonstrated at the distal radius using single-photon absorptiometry. Bone loss at this site is probably due to the predominant involvement of the hands and wrists during rheumatoid arthritis. Estimates of the frequency of generalized bone loss have varied with the measurement technique used and the study design. Studies using dual-photon absorptiometry with a radioactive source have yielded conflicting data. The most recent studies of bone mineral density used dual energy X-ray absorptiometry, which is currently the best method in terms of reproducibility and precision; results suggested bone loss in the proximal femur, whereas lumbar spine measurements were usually normal or very slightly decreased. Findings from the few longitudinal studies are discordant; this may be ascribable to differences in bone mass measurement techniques, study population characteristics, and follow-up duration. Whereas bone loss was not demonstrated in the earliest longitudinal studies, recent investigations suggested that bone mineral density was decreased in rheumatoid arthritis patients as compared with controls. Corticosteroid therapy, even in daily dosages of less than 10 mg/d prednisone-equivalent, was associated with an approximately 10% decrease in bone mass after six months, in both cross-sectional and longitudinal studies. Rheumatoid arthritis patients had a two-fold increase in fracture risk, independently from their bone mass. Factors associated with higher fracture risks were corticosteroid therapy, physical inactivity, and female gender.